Premium Bonds - Are my Bonds being missed in the draw?

Reaper

isasmurf wrote:

Of course there is no such thing as a true random number. All 'Random numbers' produced by a computer must be produced by a computer program involving some kind of algorithm and are therefore psuedo random numbers.

Not necessarily. Apparently the first Ernie machine in the 1950's based it numbers on the path taken by radioactive alpha particles, and the second in the 1970's used noise in a zener diode (no, I don't know what that is either!). You don't have to use a computer to generate random numbers.

I can't find how the current Ernie works but I'm sure that even if it uses a computer algorithm it also uses a non-computing device to seed it.

donamey

OK enough! I have decided to send the dumb letter to NSI and await patronising reply (and hopefully some winnings) - so watch out for next month's draw - I will post the results

JasonW_2

If I set up a standing order to buy bonds regularly, is there a minimum monthly amount? Hard to find it on their website. I know its £100 to start with, but its not £100 minimum a month is it?

I want to say about £50 a month then see how it goes from there, would this be OK?

Cheers
JW

Kazza242

JasonW - The minimum monthly investment by standing order is £50, according to the NS&I official website. You can find this information under the 'Our range/Premium Bonds/Investing regularly' section of the site.

Paul_Herring

Reaper wrote:

I can't find how the current Ernie works but I'm sure that even if it uses a computer algorithm it also uses a non-computing device to seed it.

For those interested. They are currently using Ernie Mark 4.

It generates random numbers using noise:

Coming up with genuinely random numbers using a computer is deceptively hard because software can only mimic true unpredictability – so eventually patterns can emerge. Software systems are therefore commonly known as "pseudo random number generators". Intel's random number generating chip uses thermal noise – changes in the voltage and heat energy given off - from transistors as its source of randomness. The thermal noise measured is a result of random electron and material behaviour in the transistors. The noise is used to control the speed of an oscillator in order to churn out binary numbers.

JasonW_2

Kazaa242 - Thanks, couldnt see this on their site. Going to invest pretty soon, quite like the idea of this way of saving, as long as I win of course! :rolleyes:

JW

isasmurf

Paul_Herring wrote:

For those interested. They are currently using Ernie Mark 4.

It generates random numbers using noise:

OK, perhaps I was wrong to say all computers use an algorithm to generate random numbers. The point is, however, the 'random numbers' are still influenced by an outside source, in this case thermal noise, and are therefore not truly random as something that happens randomly is not influenced by anything.

fabwitch_2

My Mum has a few bonds and wins about ten times a year (small amounts)but I have had bonds for 20 years and have never won anything.

webwiz

Ernie produces numbers which whilst not truly random are effectively so and the randomness is strictly monitored and controlled. I don't have any concerns in that area. However Ernie does not pick winning bond numbers because it does not know any bond numbers. The random numbers that it produces are fed into a data processing computer which has a database of bond numbers valid for the current draw, most of the random numbers do not match a bond and so are simply discarded. those that do match select the bond for a prize. This is just the sort of data processing that no other government department seems to be capable of so I would not be at all surprised if there was a bug in that part of the system. Selling part of a holding might cause the rest of the holding to be ignored. An on-line enquiry by an NSI operator might "turn them on" again. Anything is possible. One thing is sure. If anyone thinks they are not getting their fair share of prize money they should write to NSI just in case there is something in this - there is nothing to lose.

Paul_Herring

Ernie produces numbers which whilst not truly random

In what way are they not 'truly' random?